COAQNN-Based Real-Time NIFTY-50 Stock Price Prediction Using Financial Indicators and Sentiment Analysis

Navateja G1
Dr. Kiran B M2
Dr. Kaja Masthan3
Venkateshwarlu M4
Dept of CSE
Sphoorthy Engineering College
Hyderabad, India
gogulanavateja1910@gmail.com
dr.kiran@sphoorthyengg.ac.in
drkhaja.cse@gmail.com
mallikantivenkat9@gmail.com

Abstract— Stock market prediction is a complex task due to the highly volatile and non-linear nature of financial markets. Accurate forecasting of stock prices can help investors make informed financial decisions and reduce investment risks. This research proposes a COAQNN-based stock price prediction system for NIFTY-50 stocks using real-time financial data and deep learning techniques. Historical and live stock market data are collected using the Yahoo Finance API and processed using technical indicators such as RSI, EMA, MACD, volatility, moving averages, and lag features. The proposed COAQNN (Coati Optimization Algorithm–Quantum Neural Network) model predicts both mid and closing stock prices with improved accuracy. The system performance is evaluated using MAE, RMSE, MAPE, and Accuracy, and compared with Random Forest Regression and Linear Regression models. Experimental results demonstrate that the proposed COAQNN model achieved an accuracy of 98.27%, outperforming several existing approaches from the literature survey, including Logistic Regression-based prediction models with 68.62% accuracy and other hybrid forecasting methods. A Streamlit-based web application is also developed for real-time stock price prediction and analysis.

Keywords— Stock Price Prediction, COAQNN, Deep Learning, NIFTY-50, Financial Forecasting, Technical Indicators, Machine Learning, Streamlit, Yahoo Finance API